arXiv subjects

Maja H. Kirkeby

Publications and source records attributed to Maja H. Kirkeby.

13 recordsLinked to original sources

arXiv2026

Power Assumptions Matter: Evaluating End-user Laptop Energy Models for Sustainability Reporting of Browser-Based Web Services

Sustainability reporting for web-based services often relies on simplified end-user energy models that assume constant laptop power during browser interactions. Energy models such as Digst and DIMPACT apply fixed power values (15-22W), yet the validity of this approach for realistic browsing remains underexplored. We empirically evaluate constant-power assumptions in a controlled user study where ten participants repeatedly complete eight representative user flows across shopping, booking, navigation, and news services on four laptop platforms, while device energy is measured. Typical power is 9--13~W, substantially below current reporting standards, implying systematic overestimation. Moreover, the error scales proportionally with task duration, indicating systematic bias rather than random noise. Comparing progressively refined constant-power models, we find that category-specific parameters improve accuracy more than hardware-only parameters and approach flow-specific performance. The best fit is obtained by combining category (or flow) with hardware, while category-level models retain most of the benefit with fewer parameters, making them a practical upgrade for sustainability reporting.

cs.SE↗

arXiv2025

Framework Matters: Energy Efficiency of UI Automation Testing Frameworks

We examine per action energy consumption across four web user interface (UI) automation testing frameworks to determine whether consistent tendencies can guide energy-aware test design. Using a controlled client-server setup with external power metering, we repeat each UI action (refresh, click variants, checkbox, drag&drop, input-text, scroll) 35 times. Across each of the actions, energy costs vary by both framework and action. Puppeteer is the most efficient for left-click, right-click, double-click, checkbox, and input-text; Selenium is the most efficient for refresh and scroll; Nightwatch is generally the least energy efficient. The energy cost of performing the same action varied by up to a factor of six depending on the framework. This indicates that providing transparency of energy consumption for UI automation testing frameworks allows developers to make informed, energy-aware decisions when testing a specific UI action.

cs.SE↗

arXiv2025

Greening AI-enabled Systems with Software Engineering: A Research Agenda for Environmentally Sustainable AI Practices

The environmental impact of Artificial Intelligence (AI)-enabled systems is increasing rapidly, and software engineering plays a critical role in developing sustainable solutions. The "Greening AI with Software Engineering" CECAM-Lorentz workshop (no. 1358, 2025) funded by the Centre Européen de Calcul Atomique et Moléculaire and the Lorentz Center, provided an interdisciplinary forum for 29 participants, from practitioners to academics, to share knowledge, ideas, practices, and current results dedicated to advancing green software and AI research. The workshop was held February 3-7, 2025, in Lausanne, Switzerland. Through keynotes, flash talks, and collaborative discussions, participants identified and prioritized key challenges for the field. These included energy assessment and standardization, benchmarking practices, sustainability-aware architectures, runtime adaptation, empirical methodologies, and education. This report presents a research agenda emerging from the workshop, outlining open research directions and practical recommendations to guide the development of environmentally sustainable AI-enabled systems rooted in software engineering principles.

cs.SE↗

arXiv2025

Static Communication Analysis for Hardware Design

Hardware acceleration of algorithms is an effective method for improving performance in high-demand computational tasks. However, developing hardware designs for such acceleration fundamentally differs from software development, as it requires a deep understanding of the highly parallel nature of the hardware architecture. In this paper, we present a framework for the static analysis of communication within datapath architectures designed for field-programmable gate arrays (FPGAs). Our framework aims to enhance hardware design and optimization by providing insights into communication patterns within the architecture, which are essential for ensuring efficient data handling.

cs.AR↗

arXiv2024

Energy and Time Complexity for Sorting Algorithms in Java

The article investigates the relationship between time complexity and energy consumption in sorting algorithms, focusing on commonly-used algorithms implemented in Java: Bubble Sort, Counting Sort, Merge Sort, and Quick Sort. The significance of understanding this relationship is driven by the increasing energy demands of Information and Communication Technology systems and the potential for software optimization to contribute to energy efficiency. If we find a strong correlation between time complexity and energy usage, it would enhance the ability of software developers to create energy-efficient applications. This quantitative study researches the execution of four selected sorting algorithms with input varying over input sizes (25000 to 1 million) and input order types (best, worst, and random cases) on a single kernel in a Java-enabled system. The input size is adjusted according to the type's maximum execution time, resulting in 136 combinations, totalling 12960 measurements. Wall time and the CPU energy consumption is measured using Intel's RAPL. Statistical analysis are used to examine the correlations between time complexity, wall time, and energy consumption. The study finds a strong correlation between time complexity and energy consumption for the sorting algorithms tested. More than 99% of the variance in energy consumption for Counting Sort, Merge Sort, and Quick Sort depend on their time complexities. More than 94% of the variance in energy consumption for Bubble Sort depends on its time complexity. The results affirm that time complexity can serve as a reliable predictor of energy consumption in sequential sorting algorithms. This discovery could guide software developers in choosing energy-efficient algorithms by considering time complexities.

cs.SE↗

arXiv2022

Towards Comparing Performance of Algorithms in Hardware and Software

In this paper, we report on a preliminary investigation of the potential performance gain of programs implemented in field-programmable gate arrays (FPGAs) using a high-level language Chisel compared to ordinary high-level software implementations executed on general-purpose computers and small and cheap computers. FPGAs inherently support parallel evaluations, while sequential computers do not. For this preliminary investigation, we have chosen a highly parallelizable program as a case study to show an upper bound of performance gain. The purpose is to demonstrate whether or not programming FPGAs has the potential for performance optimizations of ordinary programs. We have developed and evaluated Conway's Game of Life for an FPGA, a small and cheap computer Raspberry Pi 4, and a MacBook Pro Laptop. We have compared the performance of programs over different input sizes to decide the relative increase in runtime.

cs.PF↗

arXiv2022

Energy Consumption and Performance of Heapsort in Hardware and Software

In this poster abstract we will report on a case study on implementing the Heapsort algorithm in hardware and software and comparing their time and energy consumption. Our experiment shows that the Hardware implementation is more energy efficient, but slower than the Software implementation due to a low clock frequency. It also indicate that the optimal degree of parallelization differs when optimizing for time compared to optimizing for time.

cs.SE↗

arXiv2021

An Inversion Tool for Conditional Term Rewriting Systems -- A Case Study of Ackermann Inversion

We report on an inversion tool for a class of oriented conditional constructor term rewriting systems. Four well-behaved rule inverters ranging from trivial to full, partial and semi-inverters are included. Conditional term rewriting systems are theoretically well founded and can model functional and non-functional rewrite relations. We illustrate the inversion by experiments with full and partial inversions of the Ackermann function. The case study demonstrates, among others, that polyvariant inversion and input-output set propagation can reduce the search space of the generated inverse systems.

cs.PL↗

arXiv2018

Confluence of CHR revisited: invariants and modulo equivalence

Abstract simulation of one transition system by another is introduced as a means to simulate a potentially infinite class of similar transition sequences within a single transition sequence. This is useful for proving confluence under invariants of a given system, as it may reduce the number of proof cases to consider from infinity to a finite number. The classical confluence results for Constraint Handling Rules (CHR) can be explained in this way, using CHR as a simulation of itself. Using an abstract simulation based on a ground representation, we extend these results to include confluence under invariant and modulo equivalence, which have not been done in a satisfactory way before.

cs.PL↗

arXiv2017

Confluence and Convergence in Probabilistically Terminating Reduction Systems

Convergence of an abstract reduction system (ARS) is the property that any derivation from an initial state will end in the same final state, a.k.a. normal form. We generalize this for probabilistic ARS as almost-sure convergence, meaning that the normal form is reached with probability one, even if diverging derivations may exist. We show and exemplify properties that can be used for proving almost-sure convergence of probabilistic ARS, generalizing known results from ARS.

cs.PL↗

arXiv2016

Probabilistic Resource Analysis by Program Transformation

The aim of a probabilistic resource analysis is to derive a probability distribution of possible resource usage for a program from a probability distribution of its input. We present an automated multi- phase rewriting based method to analyze programs written in a subset of C. It generates a probability distribution of the resource usage as a possibly uncomputable expression and then transforms it into a closed form expression using over-approximations. We present the technique, outline the implementation and show results from experiments with the system.

cs.PL↗

arXiv2016

On Proving Confluence Modulo Equivalence for Constraint Handling Rules

Previous results on proving confluence for Constraint Handling Rules are extended in two ways in order to allow a larger and more realistic class of CHR programs to be considered confluent. Firstly, we introduce the relaxed notion of confluence modulo equivalence into the context of CHR: while confluence for a terminating program means that all alternative derivations for a query lead to the exact same final state, confluence modulo equivalence only requires the final states to be equivalent with respect to an equivalence relation tailored for the given program. Secondly, we allow non-logical built-in predicates such as var/1 and incomplete ones such as is/2, that are ignored in previous work on confluence. To this end, a new operational semantics for CHR is developed which includes such predicates. In addition, this semantics differs from earlier approaches by its simplicity without loss of generality, and it may also be recommended for future studies of CHR. For the purely logical subset of CHR, proofs can be expressed in first-order logic, that we show is not sufficient in the present case. We have introduced a formal meta-language that allows reasoning about abstract states and derivations with meta-level restrictions that reflect the non-logical and incomplete predicates. This language represents subproofs as diagrams, which facilitates a systematic enumeration of proof cases, pointing forward to a mechanical support for such proofs.

cs.LO↗

arXiv2015

Probabilistic Output Analysis by Program Manipulation

The aim of a probabilistic output analysis is to derive a probability distribution of possible output values for a program from a probability distribution of its input. We present a method for performing static output analysis, based on program transformation techniques. It generates a probability function as a possibly uncomputable expression in an intermediate language. This program is then analyzed, transformed, and approximated. The result is a closed form expression that computes an over approximation of the output probability distribution for the program. We focus on programs where the possible input follows a known probability distribution. Tests in programs are not assumed to satisfy the Markov property of having fixed branching probabilities independently of previous history.

cs.PL↗